simplified version of https://github.com/kavgan/nlp-in-practice/blob/master/text-classification/notebooks/Text%20Classification%20with%20Logistic%20Regression.ipynb

Regression.py

import pandas as pd
import numpy as np
from sklearn.preprocessing import LabelEncoder
from sklearn.metrics import label_ranking_average_precision_score
from sklearn.feature_extraction.text import CountVectorizer, TfidfVectorizer
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LogisticRegression
from keras.utils import to_categorical

df = pd.read_json("../data/news_category_dataset.json", lines=True)

le = LabelEncoder()
df["category"] = le.fit_transform(df["category"])

cv = CountVectorizer(binary=True, max_df=0.95)
feature_set = cv.fit_transform(df["short_description"])

X_train, X_test, y_train, y_test = train_test_split(
    feature_set, df["category"], random_state=2000
)

scikit_log_reg = LogisticRegression(
    verbose=1, solver="liblinear", random_state=0, C=5, penalty="l2", max_iter=1000
)

model = scikit_log_reg.fit(X_train, y_train)

def get_top_k_predictions(model, X_test, k):
    probs = model.predict_proba(X_test)
    best_n = np.argsort(probs, axis=1)[:, -k:]
    preds = [
        [model.classes_[predicted_cat] for predicted_cat in prediction]
        for prediction in best_n
    ]
    preds = [item[::-1] for item in preds]
    return preds


top_preds = get_top_k_predictions(model, X_test, 3)
print(np.mean([actual in pred for actual, pred in zip(y_test, top_preds)]))

target = to_categorical(y_test)
y_pred = model.predict_proba(X_test)
print(label_ranking_average_precision_score(target, y_pred))

top_preds = get_top_k_predictions(model, X_test, 3)
print(np.mean([actual in pred for actual, pred in zip(y_test, top_preds)]))

model.score(X_test, y_test)